AI/ML Gives Utilities a New Tool for Wildfire Risk Mitigation

I have devoted a significant portion of my career to applying geospatial technology to solve challenges related to climate change – an issue that rises in importance each year for utilities that face increasing risks from extreme weather that drive threats like wildfires. For decades, the insights from advanced geospatial technologies have been an impactful tool for utilities seeking to mitigate risks to infrastructure, but the golden age of geospatial technology’s positive impact on risk mitigation is not behind us. It is in front of us.

That is because the insights from geospatial systems – which I refer to as Location Intelligence – are being super-charged by Artificial Intelligence and Machine Learning (AI/ML). This combination of AI/ML and advanced geospatial tools will enable utilities to address the rising risks to infrastructure and operations caused by climate change, and Exhibit A for that positive impact is how these technologies will help with wildfire risks.

To understand why, you have to start by looking at how perfectly suited AI/ML and Location Intelligence are for helping organizations perform more accurate risk analysis. Risk analysis is the single hardest task that any organization undertakes. That is why organizations dedicate so many of their most brilliant minds to it and then borrow as many external brilliant minds as possible through risk-related research and consulting. Even with all those smarts and resources, risk analysis has always been fraught with…risk. It’s time-consuming, expensive, and too often fails to accurately assess actual risks and the right mitigation steps.

Any organization doing risk analysis—regardless of industry—faces numerous variables involving its own operations and the competitive landscape. Each of those unknowns creates uncertainty that makes risk assessment difficult. But it’s even more complex for risk related to climate change. The number of variables is far more numerous, which makes assessment even more difficult.

The key to better risk assessment is to extract as many insights as possible from the massive volume of data that organizations have access to internally and externally. The volume of internal data grows exponentially from a multitude of sources: mobile devices, IoT and IT systems, customer apps, geospatial systems, infrastructure databases, and more. That is augmented by the ever-increasing volume of externally available open source data and proprietary data sets.

That data contains insights that can fill in many of the unknowns and variables I discussed above, if only organizations could extract those insights. Geospatially powered solutions have a long, successful track record of analyzing data and presenting it in visual ways that are actionable by a wide range of users. What would make it even more powerful is a way to automate that data analysis at scale, and that is where AI/ML comes in.

Let’s bring the conversation back to utilities. Mitigating the increasing frequency and severity of wildfires is one of the most important climate-related risk assessment activities of utilities today, particularly in the Western U.S. and Canada. Climate change is causing extreme weather that is creating cycles of vegetation growth and then drought conditions that are contributing to larger and more frequent fires. The implications for public safety and legal liability for utilities are enormous.

The Dixie Fire caused by an untrimmed tree limb coming in contact with a power line in northern California led to nearly a million acres burned, multiple destroyed towns, an enormous loss of life and record-breaking lawsuits. It wasn’t an isolated incident. In fact, the 15 largest wildfires have happened in this century.

In the wake of the Dixie fire and in the absence of a more sophisticated way to map this risk, many utilities have attempted to prevent fires by throwing resources at it in a financially, operationally unsustainable way. I have heard examples of utilities initially spending hundreds of millions of dollars—and in one case more than two billion dollars—on large-scale vegetation management and fire prevention initiatives since the Dixie Fire. These efforts involve small armies of field crews doing on-site inspections, supported by GIS departments analyzing as much visual data as they can process.

Throwing massive resources at this problem was an understandable reaction to the impact of those wildfires, but it was unsustainable both financially and in terms of the workload for these teams. Inevitably, those efforts have since been scaled back. The industry is clearly trying to find the right path forward in the face of these rising risks, and the answer is better risk assessment using the automation and insights of AI/ML and geospatial tools.

AI/ML and location intelligence is transforming a process that previously took months of work by that army of field crews and GIS team into an automated analysis of fire risks that only takes a few minutes. AI-driven data analysis models are able to process enormous volumes of data to identify red flag issues, enabling utilities to deploy field teams in a more focused way that uses fewer resources. These analyses can also continue in real time based on new data collected via satellite, aerial imagery, drone-captured video, sensors, cameras, and work crews—enabling utilities to adjust their mitigation efforts continuously to respond to rapidly changing circumstances.

One of the most exciting ways AI/ML will assist with wildfire mitigation is by putting these insights into the hands of more people across utilities. Natural language interfaces like ChatGPT will make it possible for professionals across a utility’s HQ as well as every worker in the field to interface directly with the underlying data to instantly get actionable information. Using natural language queries, utilities workers and contractors who do not have expertise in geospatial data analysis will be able to harness the power of those insights for their roles in the organization’s overall wildfire mitigation efforts.

Organizations have the data to mitigate the risks of wildfire. The key is to unlock that data. Geospatial technology and AI/ML are the key to doing so, and this combination of technologies gives utilities a powerful tool for protecting the public and their infrastructure.

Todd Slind is the Vice President of Technology at Locana, a TRC company, a location and mapping technology company whose software products and services solve the world’s most pressing infrastructure, sustainability, business, and social challenges. In this role, Slind leads Locana’s development of solutions that harness the power of location intelligence to support digital transformation for organizations around the world to solve business challenges as well as to support conservation, combat climate change and support humanitarian causes. He has more than two decades of experience in the geospatial industry, including his role leading Spatial Development, a custom software solution development company that used mapping technology to further sustainable social and economic development around the world. Prior to his role at Spatial Development, Slind was the Vice President of Business Development and Planning for CH2M HILL. He earned his B.A. in Systems Engineering from Evergreen State College and his M.S. in Civil Engineering from the University of Washington.